纵向数据的一种稳健同时建模方法  

A robust joint modeling approach for longitudinal data

作　　者：檀佳欣 张伟平[1] TAN Jiaxin;ZHANG Weiping(Department of Statistics and Finance University of Science and Technology of China)

机构地区：[1]中国科学技术大学统计与金融系,安徽合肥230026

出　　处：《中国科学技术大学学报》2020年第3期317-327,348,共12页JUSTC

基　　金：Supported by National Key Research&Development Plan(2016YFC0800104);National Natural Science Foundation of China(11671374,71771203,71631006)

摘　　要：使用多元t分布,提出了一种分析带有异常值的连续纵向数据的同时建模方法.不同于已有主要推断回归均值的稳健方法,本文旨在通过稳健同时参数化建模来揭示位置参数,边际尺度参数和相依参数的动态变化机制.为了加速极大似然估计过程中EM算法的速度,采用一种基于ECME的最大似然估计求解算法,所得到的估计量被证明具有相合性和渐近正态性.数据分析表明所提方法是有效的.A robust method is proposed for analyzing longitudinal continuous responses with potential outliers by using the multivariate t distribution.Unlike the existing approaches which mainly focus on the inference of regression mean,our approach aims to reveal the dynamics in the location function,marginal scale function and association by joint parsimoniously modeling the location and dependence structure.An ECME-based algorithm is applied to speed up the computation associated with the EM algorithm for maximum likelihood estimation.The resulting estimators are shown to be consistent and asymptotic normality.Numerical studies demonstrate the effectiveness of the proposed approach.

关 键 词：纵向数据 稳健估计 EM算法 同时建模 

分 类 号：O212.1[理学—概率论与数理统计]